Vibe Coding Discoverskardi
View on GitHubSkardi is a self-improving context framework for ai agents
Skardi is a Rust context framework that lets agents query data sources through SQL and MCP, records query intent, and helps turn recurring needs into reusable pipelines and tools. It supports federated queries, governed access, and RAG workflows.
Use Cases
Let agents query federated data sources through SQLRecord query intent and identify recurring agent needsPromote repeated queries into reusable pipelines and MCP toolsServe governed data context to agentsBuild hybrid search and graph RAG workflows
Built With
- Language
- Rust
- Frameworks
- Apache DataFusion · MCP
Tags
agent context · self-improving tools · MCP · SQL · query audit · data federation · RAG · pipelines · hybrid search